the same for demo and it looks the same behaviour
root@/tmp/lttng-build-root/bin # ./lttng create
Session auto-20131206-113744 created.
Traces will be written in /root/lttng-traces/auto-20131206-113744
root@/tmp/lttng-build-root/bin # ./lttng enable-event -u --all
All UST events are enabled
If:
$ gcc -o sample sample.o tp.o -ldl -llttng-ust
Fails with error: cannot find -llttng-ust, then do this instead:
$ gcc -o sample sample.o tp.o -L/usr/local/lib -ldl -llttng-ust
Daniel U. Thibault
Protection des systèmes et contremesures (PSC) | Systems Protection
Countermeasures (SPC)
As I'm using the custom image on x86 machine, i've kept the lib in
/tmp/lttng-build-root/lib folder and this sample able to see the library as
well.
And the same way I able to create the trace files for kernel but not for ust.
There were no dmesg as well.
[...]
root@/tmp # ldd sample
Yes, with that LD_PRELOAD sample able to run.
root@/tmp/lttng-build-root/bin # ./lttng create app
Session app created.
Traces will be written in /root/lttng-traces/app-20131206-174935
root@/tmp/lttng-build-root/bin # dmesg
root@/tmp/lttng-build-root/bin # ./lttng enable-event -u -a
All UST
Interesting. I've managed to sort-of-reproduce the problem. If I run
(from a user account which is NOT a tracing group member):
[...]
Then lttng correctly creates the trace folder in
/home/username/lttng-traces/a-... when sample is run, but it remains empty.
No ust sub-folder,
ping.
On 19 November 2013 10:26, Simon Marchi simon.mar...@polymtl.ca wrote:
Hello there,
liburcu does not build on the Intel Xeon Phi, because the chip is
recognized as x86_64, but lacks the {s,l,m}fence instructions found on
usual x86_64 processors. The following is taken from the Xeon Phi
Session app created.
Traces will be written in /root/lttng-traces/app-20131206-174935
root@/tmp/lttng-build-root/bin # dmesg
root@/tmp/lttng-build-root/bin # ./lttng enable-event -u -a
All UST events are enabled in channel channel0
root@/tmp/lttng-build-root/bin # ./lttng start
Tracing started
- Original Message -
From: Simon Marchi simon.mar...@polymtl.ca
To: lttng-dev@lists.lttng.org
Sent: Tuesday, November 19, 2013 4:26:06 PM
Subject: [lttng-dev] Xeon Phi memory barriers
Hello there,
Hi Simon,
While reading this reply, please keep in mind that I'm in a
mindset where
It currently fails to compile, because the Phi does not have fence
instructions.
It will use lock prefixed instructions for all memory barriers, which
is overkill for this platform, but at least it will work.
Signed-off-by: Simon Marchi simon.mar...@polymtl.ca
---
configure.ac |5 -
1
On Fri, Dec 06, 2013 at 08:15:38PM +, Mathieu Desnoyers wrote:
- Original Message -
From: Simon Marchi simon.mar...@polymtl.ca
To: lttng-dev@lists.lttng.org
Sent: Tuesday, November 19, 2013 4:26:06 PM
Subject: [lttng-dev] Xeon Phi memory barriers
Hello there,
Hi Simon,
- Original Message -
From: Simon Marchi simon.mar...@polymtl.ca
To: lttng-dev@lists.lttng.org
Sent: Friday, December 6, 2013 10:17:55 PM
Subject: [lttng-dev] [PATCH userspace-rcu] Set RCU_HAVE_FENCE to false on
Intel Xeon Phi
It currently fails to compile, because the Phi does
Can someone kindly explain the difference between system calls starting
with compat prefix and regular system calls?
Compat system calls do not appear in the system call table:
http://syscalls.kernelgrok.com/. However, I did found them in the source
code of Linux.
Thanks,
Shariyar
- Original Message -
From: Paul E. McKenney paul...@linux.vnet.ibm.com
To: Mathieu Desnoyers mathieu.desnoy...@efficios.com
Cc: Simon Marchi simon.mar...@polymtl.ca, lttng-dev@lists.lttng.org
Sent: Friday, December 6, 2013 10:40:45 PM
Subject: Re: [lttng-dev] Xeon Phi memory barriers
13 matches
Mail list logo